You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
While working on the query-scheduler documentation (PR) we realised it's quite complex to understand the implications of running a number of query-frontend / query-scheduler replicas higher than -querier.max-concurrent.
Given we assume a Grafana Mimir cluster always run with the query-frontend and given the max concurrency per querier is enforced by the number of running workers (because each worker can only execute 1 query at a time), we could simply configure the PromQL max concurrency to "unlimited".
The text was updated successfully, but these errors were encountered:
While working on the query-scheduler documentation (PR) we realised it's quite complex to understand the implications of running a number of query-frontend / query-scheduler replicas higher than
-querier.max-concurrent
.Given we assume a Grafana Mimir cluster always run with the query-frontend and given the max concurrency per querier is enforced by the number of running workers (because each worker can only execute 1 query at a time), we could simply configure the PromQL max concurrency to "unlimited".
The text was updated successfully, but these errors were encountered: